A Systematic Meta-Survey of Cyber Deception: Unified Taxonomy and Research Directions

Cyber-deception is a proactive defense paradigm that seeks to rebalance the asymmetry between attackers and defenders. Two decades of research have produced a rich but fragmented body of surveys with incompatible taxonomies, heterogeneous scopes, and uneven methodological rigor. We present the first systematic meta-survey of cyber-deception surveys to provide a unified view of the field. We group existing work into three families: general cyber-deception surveys, moving target defense surveys, and honeypot surveys. Our analysis shows that current classification models are siloed within these families and lack a shared vocabulary that spans the full spectrum of deception technologies. We also find that important techniques such as tarpits, honey encryption, and honey patches are largely absent from existing systematic surveys and their taxonomies. To address these gaps, we introduce a comprehensive taxonomy grounded in the MITRE Engage framework that integrates disparate models into a single cohesive structure. We further synthesize future research directions, highlighting the growing role of AI and machine learning in deception. Finally, we identify a persistent foundational gap: the absence of standardized evaluation frameworks and quantitative metrics for effectiveness, performance overhead, and cost benefit trade offs, which undermines claims of superiority, deployment decisions, and assessment of community progress.